Title :
A new improved CMAC neural network
Author :
Ge, Yingqi ; Luo, Xiaoping ; Du, Pengying
Abstract :
In order to accelerate the learning speed of the conventional CMAC, an Improved Credit Assigned CMAC (ICA-CMAC) is presented in this paper. And then the proposed ICA-CMAC is applied to approaching two objective functions. Simulation results show that the ICA-CMAC has faster learning speed. In addition, the paper discussed different performances of ICA-CMAC influenced by different learning rates. It was found that the ICA-CMAC with a learning rate less than 1 has a better learning performance, which might be useful in selecting an appropriate learning rate.
Keywords :
cerebellar model arithmetic computers; learning (artificial intelligence); CMAC neural network; improved credit assigned CMAC; learning speed improvement; Acceleration; Cities and towns; Convergence; Educational institutions; Electronic mail; Intelligent networks; Intelligent systems; Laboratories; Neural networks; ICA-CMAC; credibility; learning rate; online learning;
Conference_Titel :
Control and Decision Conference (CCDC), 2010 Chinese
Conference_Location :
Xuzhou
Print_ISBN :
978-1-4244-5181-4
Electronic_ISBN :
978-1-4244-5182-1
DOI :
10.1109/CCDC.2010.5498618
